#51 - historical - move to after UTA so only use if nothing else over…

…writes
SACGF · Aug 9, 2023 · 2886dfc · 2886dfc
1 parent 4759b96
commit 2886dfc
Showing 1 changed file with 26 additions and 25 deletions.
diff --git a/generate_transcript_data/refseq_transcripts_grch38.sh b/generate_transcript_data/refseq_transcripts_grch38.sh
@@ -29,6 +29,32 @@ if [[ ! -z ${UTA_TRANSCRIPTS} ]]; then
   merge_args+=(${uta_cdot_file})
 fi
 
+# Historical - these are stored in separate files for annotation/alignments
+url=https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/GCF_000001405.40-RS_2023_03/RefSeq_historical_alignments/GCF_000001405.40-RS_2023_03_genomic.gff.gz
+annotation_filename=$(basename $url)
+if [[ ! -e ${annotation_filename} ]]; then
+  wget ${url} --output-document=${annotation_filename}
+fi
+
+url=https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/GCF_000001405.40-RS_2023_03/RefSeq_historical_alignments/GCF_000001405.40-RS_2023_03_genomic.gff.gz
+alignments_filename=$(basename $url)
+if [[ ! -e ${alignments_filename} ]]; then
+  wget ${url} --output-document=${alignments_filename}
+fi
+
+filename=GCF_000001405.40-RS_2023_03_combined_annotation_alignments.gff.gz
+cdot_file=cdot-${CDOT_VERSION}.$(basename $filename .gz).json.gz
+
+if [[ ! -e ${filename} ]]; then
+  echo "Combining historical annotations and alignments..."
+  cat ${annotation_filename} ${alignments_filename} > ${filename}
+fi
+if [[ ! -e ${cdot_file} ]]; then
+  ${BASE_DIR}/cdot_json.py gff3_to_json "${filename}" --url "${url}" --genome-build=GRCh38 --output "${cdot_file}" --gene-info-json="${GENE_INFO_JSON}" --skip-missing-parents
+fi
+merge_args+=(${cdot_file})
+
+
 filename=ref_GRCh38_top_level.gff3.gz
 url=http://ftp.ncbi.nlm.nih.gov/genomes/archive/old_refseq/Homo_sapiens/ARCHIVE/ANNOTATION_RELEASE.106/GFF/${filename}
 cdot_file=cdot-${CDOT_VERSION}.$(basename $filename .gz).json.gz
@@ -123,31 +149,6 @@ if [[ ! -e ${cdot_file} ]]; then
 fi
 merge_args+=(${cdot_file})
 
-# Historical - these are stored in separate files for annotation/alignments
-url=https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/GCF_000001405.40-RS_2023_03/RefSeq_historical_alignments/GCF_000001405.40-RS_2023_03_genomic.gff.gz
-annotation_filename=$(basename $url)
-if [[ ! -e ${annotation_filename} ]]; then
-  wget ${url} --output-document=${annotation_filename}
-fi
-
-url=https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/GCF_000001405.40-RS_2023_03/RefSeq_historical_alignments/GCF_000001405.40-RS_2023_03_genomic.gff.gz
-alignments_filename=$(basename $url)
-if [[ ! -e ${alignments_filename} ]]; then
-  wget ${url} --output-document=${alignments_filename}
-fi
-
-filename=GCF_000001405.40-RS_2023_03_combined_annotation_alignments.gff.gz
-cdot_file=cdot-${CDOT_VERSION}.$(basename $filename .gz).json.gz
-
-if [[ ! -e ${filename} ]]; then
-  echo "Combining historical annotations and alignments..."
-  cat ${annotation_filename} ${alignments_filename} > ${filename}
-fi
-if [[ ! -e ${cdot_file} ]]; then
-  ${BASE_DIR}/cdot_json.py gff3_to_json "${filename}" --url "${url}" --genome-build=GRCh38 --output "${cdot_file}" --gene-info-json="${GENE_INFO_JSON}" --skip-missing-parents
-fi
-merge_args+=(${cdot_file})
-
 ## Latest
 
 filename=GCF_000001405.40_GRCh38.p14_genomic.RS_2023_03.gff.gz